The general objective of this project is to define the mechanisms by which human lymphoid cells interact with antigen-presenting cells in order to produce and regulate immune responses. Over the past year, there have been four major efforts under way that are targeted on this objective: (1) dissection of the molecular basis for peptide binding to class I HLA molecules and presentation for CD8+ T-cell recognition; (2) identification of autoreactive CD8+ T-cell responses to myelin-derived peptides; (3) identification of forms of viral peptide epitopes which can be generated in the endoplasmic reticulum and presented to CD8+ T cells; and (4) identification of peptide~class I HLA complexes recognized by human natural killer cell clones. The principal findings are as follow: (1) isolation and sequencing of endogenous peptides bound to the HLA class I molecules HLA-B14 and HLA-B44 has permitted identification of specific combinations of peptide anchor residues which can be used to successfully predict immunogenic T- cell epitopes within viral peptide sequences that are presented to CD8+ T cells; (2) peptide sequences derived from human myelin basic protein, proteolipid protein, and myelin- associated glycoprotein were identified which could bind to the HLA-A2 molecule and induce auto-reactive CD8+ cytotoxic T- lymphocyte responses in MS patients and normal individuals; (3) a viral peptide molecularly designed as a signal sequence could be cleaved from the carrier protein in the endoplasmic reticulum and bound by HLA-A2 and presented to peptide-specific CD8+ T cells on the cell surface; and (4) we are the first group to unequivocally demonstrate that human natural killer cell clones have cell surface receptors that can specifically recognize class I HLA/peptide complexes in a way which is indistinguishable from CD8+ T cells.